Performance Evaluation Of Privacy Protection Algorithm For Internet Of Vehicles Based On Location Service

With the innovation and improvement of technology,the constantly upgrading of mobile intelligent terminals and the improvement of location technology,LBS have developed rapidly,but user terminals are facing the risk of privacy leakage while enjoying convenience.A performance evaluation index and adjustable evaluation strategy of privacy protection algorithm based on LBS in compatible Internet of vehicles environment has been proposed in this thesis,to meet the performance requirements of high-speed movement of vehicles and storage burden of the system.In order to solve the problem of lacking a universal performance evaluation method for location privacy protection algorithm in the existing Internet of Vehicles environment,meet the dynamic balance between degree of privacy and quality of service,the algorithm performance indexes quantization method based on similarity model,algorithm performance evaluation method based on information entropy and privacy gain and improved adaptive performance evaluation method based on differential privacy are studied on the basis of existing researches.The specific research work is as follows:1.In order to improve the performance evaluation indexes of privacy protection algorithm in the Internet of vehicles,determine the quantitative method of indexes,the algorithm performance evaluation indexes are released from three-tier system and security framework of Internet of vehicles.Secondly,according to the measurement characteristics of distance,similarity,f-divergence and scale,the compatibility of common similarity models in the application scenarios of this thesis are compared and analyzed.Finally,the performance of the algorithm is quantified based on fusion distance,Jaccard similarity and Hellinger divergence,combined with the objective factors such as the size of anonymity set after average processing of the measurement results.2.A performance evaluation strategy DSFS-PEA(Privacy Evaluation Algorithm based on Distance,Similarity,F-divergence and Scale)for privacy protection algorithm of Internet of vehicles is proposed,which is based on the quantitative value of indexes.According to the balance between the privacy protection effect and the accuracy of the processing results,the weight based on the calculation of information entropy and privacy gain is allocated.A comparison result is drawn to substitute the above security measure model in the process of measuring the privacy gain.This thesis re-evaluates the location privacy protection algorithms derived from PPA,P2P-IS-CA-HL and SCAPGID to analyze the location privacy protection algorithms,verify and optimize the evaluation indexes and strategy.3.Using the characteristic that the query accuracy will be reduced by adding differential privacy noise,that is,adaptive selection can be made according to the application requirements of privacy protection and quality of service.Firstly,the query upper and lower thresholds of LBS are obtained,the Gini coefficient is calculated to generate a classification regression tree for privacy budget allocation.In order to facilitate the mining of improved research directions from the adaptive evaluation results of algorithm performance,it is proposed to further improve the overall performance of the privacy protection system of Io V without reducing the service quality.The proposed evaluation method is better than the traditional evaluation method in index balance and adaptivity,which provides theoretical basis and technical support for subsequent researchers to choose and improve the appropriate level of privacy protection scheme.